(Figure above from Federal Reserve Board discussion paper by Julia Lynn Coronado, Joseph P. Lupton, and Louise M. Sheiner, 2005.)
Catching up on some reading this week, and had this few-days-old post by Dylan Matthews on Ezra Klein’s blog pointed out to me. Dylan asks if the conventional economic wisdom on the “marginal propensity to consume” of richer vs. poorer people (that it’s higher for the poor mainly because they are too constrained and can’t afford to save) is really true. He points to a couple empirical studies based on the lump-sum-type rebate/refundable credit portions of the 2001 and 2003 tax cuts that suggest that the higher-income households were more inclined to spend their tax cuts than the lower-income ones were.
If these studies reflect today’s reality about how rich vs. poor would spend extra money, then this would weaken the argument that we should let the tax cuts for the rich go ahead and expire because they wouldn’t provide effective stimulus compared with tax cuts for lower- and middle-income households (and especially compared to other forms of deficit-financed stimulus).
But I have a couple of problems with leaping to such conclusions based on these earlier studies. First, the highest-income categories in these studies are much broader than the definition of “rich” being used in the debate about the Bush tax cuts–the latter being confined to households with incomes above $250,000. One thing I suspect is that households with income above $75,000 or $100,000 (the highest-income groups in these studies) don’t just have more than median incomes, but also have more than median household sizes. (There is a strong income effect in moving from middle- to upper-income levels in terms of the number of kids people have; not everyone can afford to have four children, trust me.) It is not at all surprising to me that if one does not control for family size, that households with higher incomes are more likely to spend most of their lump-sum, windfall-type tax cuts than households with lower incomes who also happen to be (at least I strongly suspect) smaller in size. If you compared the spending propensities of households with incomes over $250,000 with those of households with incomes between $100,000 and $250,000, holding constant family size, I bet we would see the latter has a higher propensity than the former.
Second, economic circumstances have really changed since the tax episodes in those studies. Everyone (the truly rich included) had been in high consumption mode at the time of the 2001 and 2003 tax cuts, because wealth effects (via housing values and investment portfolios) encouraged consumption and borrowing against that wealth. A little extra cash in the form of a tax rebate would be likely to be treated like an additional windfall, and hence spent. But the subprime mortgage and then general credit crisis shocked households into changing their borrowing and spending habits. Now households cannot count on the wealth in their homes or retirement funds as a reliable source of savings, so their propensities to consume out of any extra money (like from a tax cut) are going to be substantially lower now. So now it is not so much that the rich can just afford to save more than the poor, so that they will in fact save more of their tax cut. Today it is more that the rich can afford to pay down more of their debt (versus need to use for immediate consumption)–i.e., reduce their “negative saving.” But reducing debt or “negative saving” is the same thing as “saving more.” So I think the rich still will tend to spend a lower fraction of any non-targeted tax cut than lower-income households and that the conventional wisdom–that tax cuts for the rich aren’t great short-term stimulus (to boost aggregate demand)–still applies. (The issue of what types of tax cuts are the best for longer-term economic growth–to increase aggregate supply once the economy is at “full employment”–is a different matter entirely.)